Our units for measuring time are derived from the constancy of the earth&’s rotation, but if the planet slows down or speeds up then clocks have to take note and count one second twice so that the earth catches up, or skip one second to catch up with the earth. Although seconds, minutes and hours have their origin in the speed of the earth&’s movement, the clocks we use work on a more accurate timekeeper and, periodically, when the earth gets nearly a whole second out of step the clocks need to adjust their own count to be in time. This happens about once in two years and such an adjustment is being carried out at midnight on 30 June this year.

A number systems based on the number 60 have been found suitable for dealing with time. Just as the decimal system has become popular, the number 16 was earlier the base for weights and measures and even money — we had 16 ounces to the pound or 16 annas to the rupee, for example. Locations in computer memory, too, are best described by numbers based on 16, the Hexadecimal system. While the base 10 of the decimal system is considered easier to deal with because we have 10 digits, the bases 16 or 12, like 12 pence or 12 inches, also have their merits as the squares, 256 or 144, have more divisors than the number 100. For a simillar reason, that 60 is the smallest number that can be divided by all numbers from one to six, the sexagesimal system, or the number system based on 60, came into use in treating time.

This number system was devised by the Sumerians near modern Iraq some 3,000 years ago. The Babylonians who came after the Sumerians also used the system and in 140 AD Greek astronomer Ptolemy divided the length of the day based on 60ths. But it was around 1000 AD that Islamic scholars like Al Biruni divided the day into 24 parts and then divided the hour&’s length into 60 minutes and then 60 seconds.

The modern second is, thus, 1/(24x 60×60) = 1/86,400 of the mean solar day and, in 1874, this was adopted as the unit of time for scientific purposes. But the length of the mean solar day was found to be variable and the base was changed to the length of the day based on the time of revolution of the earth around the sun, first with reference to the constellations and then to the interval between the equinoxes. But even this definition was found inaccurate and, in 1967, the second was defined based on the frequency of radiation from the cesium atom, and this has held so far.

The system of the time of day based on the atomic clock is called the Coordinated Universal Time (known as UTC — a compromise with the French Temps Universel Coordonné), while standards of mean solar time, like Greenwich Mean Time, are known as UT1. As the second based on the atomic clock is quite close to the second based on astronomical time as well as on the average length of the solar day over a long period, UTC does not stray more than a second, over the length of a year, from UT1, and is now the standard for science and also for aviation and other civil uses.

A practical difference of UTC from the systems is that it does not observe daylight saving time. However, as the solar day is not exactly 86,400 seconds by the atomic clock, UTC does stray from solar time and needs to be corrected. The year itself, we all know, is not exactly 365 solar days but some six hours longer. We, thus, have the device to adjust the calendar, of adding a day on 29 February every fourth year as a “leap” year. In the same way, the adjustment of one second that needs to be carried out approximately once in two years to keep UTC in step with Mean Solar Time is called a “leap second”.

Keeping track of the progress of UTC and IT1, as well as announcing leap seconds, is managed by the International Earth Rotation and Reference Systems Service (known by an earlier acronym, IERS). The adjustment to be carried out on 30 June this year would be that the UTC clock, at midnight, would mark time for one second to slow down to the time kept by the rotation of the earth. Thus, where the clock should have moved from 11.59.59 to 12.00.00, it would first move to 11.59.60 and then to 12.00.00. This is just like February not moving from the 28th to 1st March, but whiling a day away, as the 29th.

Unlike the adjustment for a leap year, which is always to add one day to the calendar, the leap second may be added or taken away. The reason is that the leap second correction is not only for a constant difference factor but for both the increase of reduction of the length of the day due to climatic changes or geological occurrences or even ocean currents, or tides, in the ocean or within the land mass.

East-west movement in the atmosphere would be compensated by corresponding changes in the rotation of the earth. Warming or cooling, which would cause the atmosphere and also the oceans to expand or contract would also have the effect of a spinning ballerina stretching her arms out or drawing them in, and would slow or speed up the earth&’s rotation.

The 25 corrections since the system of correction started in 1972, however, have all been cases of delaying UTC by one second and none to save a second. This is to say that the solar year has been a little slower than the year according to the atomic clock. Apart from the way the atomic second is defined, it is also a fact that earth&’s rotation is slowing by a barely perceptible extent for centuries now.

The effects of atmospheric and geological changes have partly compensated, but not in a uniform way. The difference between the two clocks is, hence, not regular and the time taken for the difference between the two to grow to one second has not been the same every time it happens. To have 25 corrections in 43 years works out to one correction in less than every two years. But the last correction was in 2012 and we can see that the 26th, on 30 June, will be well over two years later. In fact, there were no corrections between January 1999 and December 2005, but there were nine corrections in the eight years from 1972 to 1979.